Roadmap for focused ion beam (FIB) technologies

Executive Summary

A focused ion beam (FIB) instrument employs a finely focused beam of ions with an energy of typically 2 keV to 30 keV that is scanned across a sample to effect modifications down to the nanometer scale. The technique relies on the transfer of energy from the energetic primary ion to target atoms during elastic nuclear collisions, resulting in the displacement of these substrate atoms and potentially their removal from the sample. As such the FIB instrument is a universal tool for all areas of research enabling comprehensive analyses, maskless local material changes and rapid prototyping of devices.
The aim of this document is to provide an overview of the current state of the art of FIB technology, its applications and important tool developments, all of which require attention by researchers and technologists developing new FIB-based workflows and instrumentation. As such, this document can serve as an important reference work for students, FIB users, academic and commercial developers of related technologies, and funding agencies. It includes overview tables, providing a bird’s eye view of the relevant works in this field. New developments in the various driving fields of research pose new challenges for FIB techniques, for which we propose targeted solutions and provide a list of technical developments that will be required. 

From an instrumentation point of view, the ubiquitous gallium FIB is an excellent tool for the preparation of both transmission electron microscope (TEM) samples and functional nanostructures. However, FIB sources producing ions of other elements are becoming increasingly common and provide additional degrees of freedom for the fabrication of new meta- and functional materials. In fact, liquid metal alloy ion sources (LMAISs) and gas field-ionization sources (GFISs) were used even before the gallium liquid metal ion source (LMIS)  started to control the market. Newer developments include the plasma focused ion beam (PFIB) (using electron cyclotron resonance (ECR) or inductively coupled plasma (ICP) sources), which has quickly gained relevance due
to its ability to remove large volumes of material while maintaining a lateral resolution sufficient for many applications. Combined with the very recent addition of the magneto-optical trap ion source (MOTIS), about two thirds of the elements of the periodic table can now be accessed for FIB applications (see Fig. 1 and Tab. I). New applications related to, e.g. health, quantum technology, as well as energy conversion and storage, present new challenges to existing techniques that will require innovations in ion sources, beam transport and detectors (see Fig. 31). Low energy FIBs (≤1 keV) are particularly relevant for the semiconductor industry, quantum technology, and applications based on low dimensional materials.

New developments related to heavy and unconventional ion beams are important for single ion implantation (SII) and analytical applications, such as secondary ion mass spectrometry (SIMS). The theory of ion-solid interaction and all related processes that may occur in FIB processing must bridge many orders of magnitude in terms of time and length scales (see Fig. 5). Therefore, a tradeoff must always be made between accuracy and modeled system size. The binary collision approximation provides sufficient insight into interesting FIB-related quantities that are mainly driven by the ballistic phase of the collision cascade, such as sputtering yield, ion penetration depth, and degree of amorphization (see Tab. III for an overview of available binary collision approximation (BCA) codes). In addition, other techniques such as molecular dynamics (MD), kinetic Monte Carlo techniques (kMC), density functional theory (DFT), and continuum modeling can address further important aspects such as short-term defect evolution (MD); long-term system evolution at elevated temperatures (kMC); charge state of defects and charge transfer (DFT); and ion-induced chemistry as well as surface processes (continuum modeling).

An important application technique in this respect is gas-assisted processing, usually referred to as focused ion beam-induced deposition (FIBID). This is the nanometer-scale counterpart to industrial and consumer 3D printing. The variety of precursor gases (see Tab. IX) and the patterning flexibility of the FIB enable the fabrication of nanoscale devices with special properties and geometries for emerging applications (see Tab. VIII). Here, an in-depth understanding of the complex processes related to precursor gas flow, surface diffusion of molecular species, and modeling of the generation of secondary particles are needed for accurate process simulation.

Process modeling is also relevant for other FIB applications, such as subtractive processing (see Tab. V) and defect engineering (see Tab. VI). Applicability, however, will also require an extension of the existing knowledge on ion-solid interactions into the energy range relevant for FIB applications. This knowledge transfer would be particularly helpful for calculations of the production rate of charged secondary particles, especially secondary electrons (SEs). The yield of SEs is relevant for several applications, including SII and also general imaging and analysis. For analysis applications, robust models for the prediction of secondary ion (SI) yields are needed.

In addition to addressing these fundamental questions, efforts are required in the areas of (empirical) process modelling and shape prediction for additive and subtractive patterning. Here, open application programming interfaces (APIs) and open source software in combination with machine learning (ML) will enable better control over the focused ion beams, which will in turn allow higher precision, resulting in faster and more reproducible device fabrication. New software and computing technologies also enable the creation and evaluation of large data sets that are mostly generated via FIB-based analytical methods, e.g. FIB-SIMS and serial sectioning (see Tab. VII). ML algorithms can aid with the acquisition, alignment, and segmentation of these typically multidimensional data sets. Drivers for these techniques are health, energy storage and conversion, meta- and functional materials, and the semiconductor industry. Currently, biological applications are also driving this field through the use of FIBbased tomography to generate nanoscale 3D representations of samples, enabling new insight into the microscopic structures responsible for cognitive processes.

As the applications of FIB processing are extremely diverse, it will not be a single technological development that brings a breakthrough. Instead, numerous advances will be needed in order to address the challenges that have been identified and the novel developments will often impact several scientific fields, as can be seen from the overview graph in Fig. 28. Therefore, we look to an exciting future where FIB processing will grow in importance and open new avenues in many areas of science and technology, spanning quantum technology to the life sciences.